The mechanisms by which drugs affect human behavior are complex, involving the interaction between the direct actions of the drug (e.g., impaired coordination) and the functional behavioral effects such as altered motivation. Specific determinants of drug response include the drug dose, the route of administration, the person's physiological and psychological state, the particular environmental conditions, and the nature of the behavior or test being measured. One approach that we are pursuing is to vary the environmental conditions by mainpulating the reinforcement (monetary) contingencies under which subjects perform various tasks. This can be thought of as manipulating a subject's motivation to perform. By also varying drug dose and using tests that measure different aspects of performance (e.g., psychomotor vs. cognitive), we can begin to explore the complex interactions underlying the effects of drugs on behavior. In one such study, the experimental design was a factorial crossing of three marijuana doses with two reinforcement contingency conditions. In one condition, subjects earned $.01 for every three correct responses on the performance tests, and in the other, they earned $.15 per three correct responses. Testing occurred at predrug baseline and 5 min and 2 hr postdrug (word recall) and 5, 45 min, 2, 4, 10 hr postdrug (psychomotor tests). High dose marijuana decreased the number of correctly recalled words in the delayed free recall test. There was a significant dose X reinforcement contingency interaction with respect to intrusion errors on the memory test. Under the low contingency condition, intrusion errors increased in a dose-dependent manner. However, under the high contingency condition, intrusion errors showed an inverted-U dose function, such that errors were reduced to placebo levels after high dose marijuana. These data indicate that marijuana-induced impairment in reflective cognition can be influenced by reinforcement contingency. Another approach to investigating the mechanisms by which drugs influence behavior is to focus on performance impairment produced by psychoactive drugs. The performance-impairing effects of drugs of abuse produce a large toll on the nation each year in terms of traffic injuries and fatalities and lost productivity in the workplace. However, for most drugs, we lack basic knowledge about the behavioral mechanisms underlying their impairment of human performance. A series of studies is being conducted that will address such questions. A battery of physiological, behavioral, and performance measures designed to determine whether an individual is behaviorally impaired as the result of taking a drug has been tested with ethanol, marijuana, and cocaine. A second study investigated the effects of marijuana, amphetamine, codeine, and alprazolam on this same test battery as well as a cognitive test of attention and memory. By also collecting blood samples during these studies, we are able to gain valuable information concerning the relationship between plasma concentrations of drugs and degree of performance impairment. In two studies, we investigated the effect of smoked marijuana on four standardized field sobriety tests (FST) that are used to determine whether a person can safely operate a motor vehicle. Subjective effects and delta-9-tetrahydrocannabinol (THC) plasma concentrations were also measured to correlate with behavioral impairment. In a residential study, 12 volunteers participated in six experimental sessions. At each session, subjects smoked ad lib two half-cigarettes containing 0 or 3.58% THC. Placebo, low, and high doses consisted of two placebo half-cigarettes, one placebo and one active half-cigarette, and two active half-cigarettes, respectively. Subjects received each marijuana dose twice in random order. Marijuana impaired performance on only one FST, the One Leg Stand, by increasing number of hops and times the elevated foot touched the floor to maintain balance. In a nonresidential study, 20 subjects participated in three experimental sessions. At each session, subjects smoked two cigarettes (16 paced puffs) containing 0, 1.75, or 3.55% THC. Marijuana impaired performance on two FST, One Leg Stand and Finger to Nose. The number of times subjects put their foot down and raised their arms to maintain balance and amount of body sway were increased by marijuana in the One Leg Stand test. A dose-dependent increase in number of misses was observed in the Finger to Nose test. In both studies, marijuana produced orderly dose-related increases in subjective ratings of intoxication. THC plasma concentrations peaked immediately after smoking and had declined to 15-28 ng/ml at time of FST testing (15 min postsmoking). These data suggest a threshold plasma THC level in the 20-25 ng/ml range for marijuana to impair behaviors critical for safe driving.